Sunday, October 3, 2010

Let's do more computer science

Let's do more computer science: [...] But I'd like to see young computer science students get up and present to their teachers some new computer science. [...] Example: Here's a group of researchers at ten universities scattered around the world. They're collaborating on a presentation they will do at a conference in eight months. Most of the research is done, they just have to assemble it in a convincing way, while giving each contributor proper credit for their work. Here's a list of the current tools they are using. At this hackathon each group will create a tool that in some new way helps them collaborate. The projects will be judged strictly on utility, not eye-candy appeal. [...] Don't worry though, eventually there will be money here. But by putting it first with people who are better at using the Internet than pitching venture capitalists, we all get a chance to have some fun, learn from each other, and actually accomplish something useful in the time we have together.
I really like this idea of Dave Winer's. His collaboration tool example is a favorite of mine. I've faced the problem many times, and none of tools around are even close to adequate for the task. We have either cloud-based or P2P doc systems that impose their common denominator document formats, or, at the other end of the spectrum, some version control system where we store document files plus some conventions to avoid stepping on each other while editing ("You have the token for Section 2").
As to Dave Winer's main point, he's right on the money. Current computer science education has too few opportunities for extended focused creativity. Class projects are most often dead-end reconstructions of old solutions for established problems rather than new solutions for unsolved problems. In part this is because school time is fragmented into many small slices, in part because organizing teaching and student evaluation around open-ended projects that are likely to fail (as most experiments do) is really hard. It's interesting to note that those two difficulties, fragmented time and the likelihood of failure, are also very serious issues within companies. We need to come up with better ways of framing collective learning, discovery and invention that encourage creativity, get value even from failure, and give fair credit to individual contributions.

No comments: